Kibble-Zurek problem: Universality and the scaling limit